whatmodel.sh / Phi-4-mini-reasoning 3.8B

Phi-4-mini-reasoning 3.8B

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Phi-4-mini-reasoning 3.8B is an open-weights language model from Microsoft, released April 2025. Its published size is 3.8B parameters and a 128k-token context window. Licence: mit.

Model information sources

· Parameter count ↗

Also: "the 3.8B parameters Phi-4-mini-reasoning model". Repo safetensors header total 3,836,021,760 (HF API); sibling phi-4-mini entry uses 3.84. Dense: active_b = params_b.

47.0±2.2#397 of 719 LLMsfrom 6 benchmark results across 3 sources

$ how it compares

The LLMs ranked either side of Phi-4-mini-reasoning 3.8B, out of 828 with a score, then more from Microsoft.

#modelscorewhereRAMreleased
472GLM-4 9B Chat Zhipu AI (Z.ai)47±2local~7.9GB2024-06
473Yi-1.5 34B (base) 01.AI47±3local~22GB2024-05
474Phi-4-mini-reasoning 3.8B Microsoft47±2api $0.13/M +local~3.8GB2025-04
475Qwen1.5 110B Chat Alibaba47±1local~45GB2024-04
476WhiteRabbitNeo-33B-v1 WhiteRabbitNeo47±1local~22GB2023-12
more from Microsoft
252Phi-4-reasoning-plus 14B Microsoft59±1local~10GB2025-04
261Phi-4-reasoning 14B Microsoft58±1api $0.22/M +local~10GB2025-04
365Phi-4 14B Microsoft53±1api $0.087/M +local~10GB2024-12

see all LLMs · find the ones that fit your machine

$ what it's good at

math▁▂▃▅▆▅▄█▄▁▁▁ #178 of 398 * · 32% of its score · 2 tests

Beats MiniMax-M2.1, GPT-4.1, DeepSeek V3 0324 here, though they rank higher overall.

testresultvs the others in its rundifficultycountswhy
AIME 2024 lab57.5% ✎▁▁▁▂▁▃▁▅███▃ beats 27/85medium · avg 53%▇▇▇▇▇▇▇▇▇▇ 25%sharp · at its level
MATH-500 lab94.6% ✎▁▃▂▃▅▇▅▇█▅▂▁ beats 52/84very easy · avg 87%▇▇▇ 6.8%noisy
other tests4 not in a task · 66% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond lab52.0% ✎▁▁▁▁▂▄▅▅███▄ beats 93/310hard · avg 51%▇▇▇▇▇▇▇▇▇▇▇▇▇ 32%sharp · at its level
BRIDGE clinical text, zero-shot average (87 tasks)21.3% ✎▁▁▁▁▁▃▂▃▄▆█▃ beats 14/97hard · avg 32%▇▇▇▇▇▇▇▇ 20%blunt · at its level
EuroEval HellaSwag (English), test split, 5-shot generative MC, accuracy41.0% ✎▃·▁▁▂▃▂▅█▇▄▁ beats 22/100medium · avg 68%▇▇▇ 7.7%at its level · noisy
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy55.7% ✎▁▁▁▁▁▁▁▃▄█▄▃ beats 26/146easy · avg 83%▇▇ 5.5%at its level · noisy

$ where to run it

Through an API

api · Microsoft Azure · $0.13/M · $0.075 in / $0.3 out per 1M tokens · 128k context · azure.microsoft.com ↗ · price sourceUSD per 1M tokens, Azure AI Foundry Models pay-as-you-go (Microsoft models table), 'Phi-4-mini-reasoning' (GA), 128K context: input $0.075, output $0.30, region US East. Currency: USD. Microsoft Azure official pricing page (price cells carry the amounts in the page's own data), observed 2026-09-25.

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Phi-4-mini-reasoning 3.8B · needs ~2.49 GB · from huggingface.co/unsloth
    chat in the terminal
    llama cli -hf unsloth/Phi-4-mini-reasoning-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf unsloth/Phi-4-mini-reasoning-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Phi-4-mini-reasoning 3.8B · needs ~2.49 GB · from huggingface.co/unsloth

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/unsloth/Phi-4-mini-reasoning-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Phi-4-mini-reasoning 3.8B · needs ~3.15 GB · from ollama.com
    chat in the terminal
    ollama run phi4-mini-reasoning:3.8b

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Phi-4-mini-reasoning 3.8B · needs ~2.2 GB, Mac only · from huggingface.co/lmstudio-community
    chat in the terminal
    mlx_lm.chat --model lmstudio-community/Phi-4-mini-reasoning-MLX-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model lmstudio-community/Phi-4-mini-reasoning-MLX-4bit

Official weights and documentation: huggingface.co/microsoft ↗

$ how the score was worked out · microsoft/Phi-4-mini-reasonin… 64%, BRIDGE Medical Leaderboard (H… 20%, EuroEval leaderboards (result… 13%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
microsoft/Phi-4-mini-reasoning model card · lab-reported3▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 64%MIT
BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)1▇▇▇▇▇▇ 20%MIT
EuroEval leaderboards (results collection), English generative leaderboard2▇▇▇▇ 13%MIT
prior · pull toward the middle▇ 2%
weighted average= 47.0

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json